“Follow-The-Workload” Beats the Latency-Survivability Tradeoff in CockroachDB

Re:“Follow-The-Workload” Beats the Latency-Survivability Tra... by arrty88 on Tuesday December 05, 2017 @08:10PM (#15857708)

We had it with RethinkDB already!

[ Parent | on HN ]

Re:“Follow-The-Workload” Beats the Latency-Survivability Tra... by manigandham on Tuesday December 05, 2017 @09:08PM (#15857940)

Except it has lacking performance and is now impractical without a backing support company.

[ Parent | on HN ]

Re:“Follow-The-Workload” Beats the Latency-Survivability Tra... by a012 on Tuesday December 05, 2017 @08:54PM (#15857888)

Does rethinkdb have ACID? No?

[ Parent | on HN ]

Re:“Follow-The-Workload” Beats the Latency-Survivability Tra... by elvinyung on Tuesday December 05, 2017 @09:00PM (#15857911)

Whether that actually matters is arguable: No one has performant distributed (i.e. cross-partition) ACID. (Except maybe Google.)

For the thing that does matter, which is ACID within the same partition/entity group, I would argue that RethinkDB already does have that.

Mike Stonebraker says that there are no fast distributed transactions [1]. I'm inclined to believe him, and that's not just an appeal to authority.

[ Parent | on HN ]